Increasing stimulus similarity drives nonmonotonic representational change in hippocampus
- Department of Psychology, Yale University, United States
- Department of Psychology, Queen’s University, Canada
- Department of Psychology, Princeton University, United States
- Princeton Neuroscience Institute, Princeton University, United States
Figures
Figure 1
Explanation of why moderate levels of visual similarity lead to differentiation.
Inset (bottom left) depicts the hypothesized nonmonotonic relationship between coactivation of memories and representational change from pre- to post-learning in the hippocampus. Low coactivation …
Figure 2 with 2 supplements
Schematic of image synthesis algorithm, fMRI task design, and behavioral validation.
(A) Our image synthesis algorithm starts with two visual noise arrays that are updated through many iterations (only three are depicted here: i, ii, and iii), until the feature activations from …
Figure 2—figure supplement 1
Schematic of synthesis algorithm, related to Figure 2A.
(A) In the endpoint selection phase, images were generated that maximally express each of the feature channels in a later layer of a deep neural network. These feature activations were then …
Figure 2—figure supplement 2
Model validation, related to Figure 2A,B.
To ensure that our model-based synthesis approach was effective in constraining the shared features among image pairs, we fed the final image pairs (Figure 1B) back through the neural network that …
Figure 3 with 2 supplements
Analysis of where in the brain representational similarity tracked model similarity, prior to statistical learning.
(A) Correlation of voxel activity patterns evoked by pairs of stimuli (before statistical learning) in different brain regions of interest, as a function of model similarity level (i.e. how similar …
Figure 3—figure supplement 1
Similarity tracked model similarity prior to statistical learning, related to Figure 3.
Correlation of voxel activity patterns evoked by pairs of stimuli (before statistical learning) in different brain regions of interest, as a function of model similarity level (i.e. how similar the …
Figure 3—figure supplement 2
Analyses of medial temporal cortex subregions, showing neural similarity (prior to statistical learning) as a function of model similarity, related to Figure 3A.
Correlation of voxel activity patterns evoked by pairs of stimuli (before statistical learning) in different brain regions of interest, as a function of model similarity level (i.e. how similar the …
Figure 4 with 2 supplements
Analysis of representational change predicted by the nonmonotonic plasticity hypothesis.
(A) Difference in correlation of voxel activity patterns between paired images after minus before learning at each model similarity level, in the whole hippocampus (HC) and in hippocampal subfields …
Figure 4—figure supplement 1
Analyses of medial temporal cortex subregions, showing representational change as a function of model similarity, related to Figure 4A.
Difference in correlation of voxel activity patterns between paired images after minus before learning at each model similarity level. Model fit was not reliable in PHC (r = 0.076, 95% CI = [−0.055 …
Figure 4—figure supplement 2
Exploratory analyses of the association between visual similarity effects in visual regions of interest, and the observed nonmonotonic effect in DG, related to Figure 4A.
(A) For each participant, we extracted the statistic for the correlation between model similarity level and representational similarity in visual regions of interest (Linear Effect), as well as the …
